Big Data Technology, Streaming Music, And Pandora's Competitive 'Moat'

Jul. 2.14 | About: Pandora Media (P)


Neither Apple's purchase of Beats nor Amazon's recent music launch represent direct competition to Pandora which competes for revenue with terrestrial radio broadcasters (e.g. Clear Channel, Cumulus & CBS).

Pandora is really a Big Data technology company which uses its industry-leading data analysis algorithms to create personalized music playlists.

Its big data technology distinguishes Pandora from the other music streaming/internet radio services and creates a competitive 'moat' that neither Google, nor Apple, nor Amazon (much less Spotify) can challenge.

Undervalued & Misunderstood, Pandora Offers Long-Term Growth & Value at Current Levels

The music business has been transformed over the last two decades by the emergence of Napster, the evolution of digital music downloads and more recently by the proliferation of real-time 'streaming' music services which have destroyed traditional notions of "ownership" and the commercial transactions, which accompanied them. While technology continues to revolutionize the manner in which individuals access music, listeners' basic consumption patterns and preferences have remained largely unaltered with some people desiring an interactive, on-demand experience in which they pick-and-choose exactly what they want to hear and when they want to hear it, and others preferring a more-or-less personalized but ultimately still passive, radio-like experience. While both interactive, on-demand music services (e.g. Spotify, Rdio & Beats Music) and non-interactive radio-like providers (e.g. Pandora (NYSE:P), iTunes Radio (NASDAQ:AAPL)) are now available through streaming music services--and notwithstanding the fact that people increasingly conflate the two different categories--they are in fact, very different businesses, offering different products, competing for different (though perhaps overlapping) listeners and employing wholly distinct business models.

My focus in this note is on the passive, radio-like streaming music business generally and Pandora specifically, which dominates this category of the internet music space with nearly 80% share. Despite multiple competitors' best efforts to market and promote alternative services, neither Clear channel's iHeartRadio, nor Apple's iTunes Radio (a passive complement to their original iTunes product), nor any other would-be competitor have been able to take share. In fact, quite the contrary, Pandora's share has actually increased since iHeart's and iTunes Radio's respective launches. Moreover, while Pandora indirectly competes with on demand services, like Spotify, Rdio & Beats for time spent listening to music, none of these companies attempt to sell local advertising or otherwise, constitute a direct competitor to Pandora's business. This is an important, and seemingly, little understood fact as there are seemingly numerous articles every week about how Pandora is surrounded by aspirants to its leadership spot and how deep-pocketed competitors are going to undermine their business. What these analyses all fail to recognize however, is that Pandora is really a big-data software company that utilizes its proprietary technology to develop algorithms that take all of the assorted information they possess on the music they license and their listeners' tastes and preferences, and then use that data to create individualized radio stations playing songs tailored to a users' specific tastes. This is critical because it is this proprietary technology that not only differentiate Pandora from the myriad of other firms in the music streaming space, but also creates a competitive 'moat' around their business that can't be penetrated just because a large company with extensive resources (e.g. Apple, Google, Amazon etc.,) and an ability to license the same catalog of music decides that it wants to offer its customers an internet streaming music product. While that would-be competitor can perhaps threaten Spotify or Rdio or one of the other interactive, on-demand streaming services which effectively offer their customers a virtual version of the now bankrupt music superstores (i.e. a place to purchase/access pre-recorded music), they can't readily duplicate Pandora's Music Genome Project data much less its algorithms for taking that data alongside data on its individual customers and generating custom playlists--something Apple has attempted to do for years with iTunes without success.

Rather than the other streaming music services, Pandora competes for revenue against terrestrial radio broadcasters, the largest of which are Clear Channel, CBS (NYSE:CBS) and Cumulus (NASDAQ:CMLS). Most market estimates project the terrestrial radio market to be a $15-17bn market, 70-80% of which is local. This is why Pandora is building out a local ad sales force, to compete for this revenue. Pandora's share of radio audience is now up to 9%, and growing. More significantly, 9% likely understates Pandora's true share of the market because nearly 50% of radio listening takes place in the car, and Pandora is just starting to get into that side of the market. Thus, Pandora has 9% share of the total terrestrial radio market while addressing only 50% of the market--put differently, Pandora has captured nearly 20% of the terrestrial radio market in which it competed! Additionally, the firm is showing rapid progress in the automobile market: Pandora is already in 10 of the top 10 selling car models and has been installed in one-third of all new cars sold in the US last year. Of those nearly 5m cars, we estimate Pandora was activated in over 80% of them. Pandora has built a large audience that will get much larger.

Over the past year there are several developments that will enable Pandora to take revenue share from terrestrial radio:

1) Pandora is now integrated with Telmar's ad planning software;

2) Pandora is now integrated with Strata and Media Ocean's ad buying software;

3) Pandora's audience share is measured by Triton and is MRC certified in 275 local markets (Arbitron is only MRC certified in approximately 30 markets).

As Pandora takes audience and revenue share away from terrestrial radio it is very difficult for traditional radio players to respond, because Pandora is leveraging the internet's strengths via personalization and interactivity in a way that terrestrial radio cannot. Furthermore, the largest terrestrial competitor, Clear Channel is a bankruptcy waiting to happen… last year Clear Channel had approximately $6bn of revenue, operating profit of $1.2bn and interest on $20bn of debt equal to $1.6bn (or 133% of operating profit). Bottom line, Clear Channel is hemorrhaging capital and with debt maturities over the next couple of years depriving them of any financial flexibility, they will be hard-pressed to compete.

Pandora's q1 results were strong. Mobile revenue was up 92% year over year. Local revenue was up 232% year over year… 232%. Local revenue is now 20% of total revenue. Pandora's q1 gross margins expanded 18% year over year to 32% (gross margins have expanded year over year in every quarter since q1 2013, that's five in a row and we expect the trend to continue). Pandora reinvests its gross profit in sales hires and engineers to drive future growth in sales and product innovation. As Pandora builds out its local sales force its revenue can increase dramatically, without increasing ad load. Local ads earn close to 4x what national ad CPMs earn. With only 20% of Pandora's revenue from local ads it leaves a lot of upside as local grows to approach the 70% level typical for radio businesses. We expect Pandora's ad revenue growth to re-accelerate in the second half of 2014 as sales reps hired during q1 begin to contribute to revenue.

Many commentators criticize Pandora's business model for lacking a competitive advantage-they miss the point. What Pandora does, like Google, seems easy but is actually quite hard-switching costs seem low, but no one's switching. Like Google vs. Bing in search (and Google vs. Apple in maps), Pandora provides a much better experience than its competition. The moat that enables this advantage is large and defensible-Pandora simply has better technology which enables them to take the data they possess about how consumers want to listen to radio in a non-interactive environment--over 35 billion thumbs up/down and growing--and about the music in their catalog and transform that information into personalized playlists that no other company can come close to replicating. This is why people who sample iTunes Radio switch back to Pandora… iTunes Radio is not nearly as good a product as Pandora, not because iTunes lacks information on what music their clients/customers like and have purchased in the past, but because they can't take that data, combine it with other information about the individual and cross-reference those data points against a catalog of music to put together playlists that don't simply regurgitate what a customer has previously purchased but also offer new selections specifically chosen because they are likely to resonate with different aspects of that person's musical taste and personal habits and preferences. This is Pandora's 'secret sauce' and it is not something that Apple, Amazon, Google, much less Spotify, Rdio etc, have been able to replicate and 77 million monthly active users know it. The result is, notwithstanding all of these would-be-competitors, not only is Pandora's audience size at an all-time high, but its engagement, at over 22 hours per listener per month, is as well. Apple, Google, Amazon and others can all attempt to aggregate data about what people listen to and have purchased and this may be enough in an interactive (Spotify-like) environment, but it is barely the beginning, and on its own virtually meaningless, in a non-interactive (Pandora-like) environment if you don't also have the technology to manage that data.

Another attack on Pandora is that it's not making any money, therefore the current PE multiple is unsustainable and too high. It is unsustainable, but it is not too high… and in the near term it should be much higher. Like a building under construction in a prime location, which is currently beset by high production costs and little-to-no incoming tenant revenue, Pandora is a high quality asset in the process of investing in its expanding local advertising footprint. While in both instances, the short-term outcome is greater losses, those current investments will, longer-term yield positive returns and like with an unfinished building in a tier 1 city, profitability is just a matter of time and the asset has significant value notwithstanding its current pre-profit state. Indeed, as Pandora's share reaches critical mass, to attract local advertisers, the company is scaling a local ad sales force (completing the construction and high-end finishes in the building analogy). As Pandora's share continues to climb, its ability to generate earnings will likewise increase. Most of the Street understates Pandora's share of radio 3-5 years out. By penetrating the car, engagement with Pandora will increase and hours per user per month will be materially higher than Street models currently predict. So investors positing a negative or short thesis on Pandora will need to articulate why the company won't be successful hiring local sales reps to sell radio-like ad inventory-because unless they fail to do that, we believe they are going to generate significant earnings over the next couple of years.

Furthermore, it is highly likely that the Copyright Royalty Board decreases Pandora's content cost burden. The company will not discuss this, but they do not have to-the facts are on their side. Pandora pays a per track fee per song (or 25% of revenue-if the per track fee amounts to less than 25% of revenue), historically this per track fee has amounted to over 50% of revenue, while terrestrial radio pays zero and satellite, under the 801b standard, pays approximately 10%. Canada just completed a process similar to the CRB; it was a multi-year process that culminated in 10 days of testimony. Canada is a very important data point because the arguments made in Canada represent a snap shot of what will be argued before the CRB in the US between now and December 2015. Canada's ruling resulted in a per track rate equal to 5% of revenue. That's 5% of revenue! If the CRB were to rule similarly for 5% in the US, nearly 50% of Pandora's revenue would immediately drop to their bottom-line-that's $2 of earnings this year and over $5 in 3 years (we expect Pandora's revenue to exceed $2.7 billion in 2018). Even if the CRB ruled at 25% of revenue (the maximum possible given numerous data points) it still is materially accretive to most Street estimates. A 5% is also not a ridiculous number for sound recording rights: publishers get 3-4% of revenue for publishing performance rights and it is difficult to argue that sound recording is worth more than the publishing right (the publishing right might be too low, but that is a different point).

We have not discussed international expansion, but obviously Pandora's TAM will grow over time. Pandora has made great progress in Australia and New Zealand. We believe expansion into Canada is going to happen during the next 12 months. Pandora has recently made new hires to focus on international expansion and has said their own internal projections indicate they would have over 1 billion registered users (they have 250 Million, currently) had they not blocked non-US users in 2007.

Most importantly, Pandora is led by a very strong management team. Brian McAndrews as CEO and Mike Herring as CFO have created value for shareholders at prior companies, both maintain excellent reputations per former colleagues and prior venture backers. McAndrews recently hired Sara Clemens to lead Pandora's international expansion.

Pandora is currently valued at approximately 5x 2015 sales, and based on our projections fair value is $56 at the end of 2014. Our 2018 revenue is $2.634 billion, EBITDA is $850 million and EPS $2.48 (holding content costs constant with 2015 rate). If we imagine that the CRB cuts their content cost to 25% of revenue (still 5x the Canadian court's recent) EBITDA jumps to 1.192 billion and EPS to $3.50.

Disclosure: The author is long P. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it. The author has no business relationship with any company whose stock is mentioned in this article.